Portfolio Insurance andStock Market Risk

Concern about stock market volatility hasreached a record high. This spring, John J.Phelan Jr., chairman of the New York Stock Ex-change (NYSE), launched a personal crusadeagainst program trading, warning all who wouldlisten that the sophisticated trading strategiesnow in use-as distinct from changes in funda-mental values-could lead to a "meltdown" ofthe stock market. Phelan joined the chorus ofcomplaints led by Representative John D. Din-gell, chairman of the House Energy and Com-merce Committee, who charges that programtrading "is rapidly destabilizing our capital mar-kets and eroding investor confidence."

In the burst of activity surrounding thesecharges, the NYSE recently announced the ap-pointment of former attorney general NicholasKatzenbach to head a study of program trading.In addition, Phelan has proposed increasingmargin requirements on stock index futures con-tracts and, in the event of a drop in the stockmarket of 25 percent or more, imposing tradinghalts on selected stocks.

With this image of the Great Crash nowfirmly in mind, it is worth taking a look at whatlies behind the meltdown scenario. To beginwith, the alleged culprits, referred to by Phelanand the popular press as program traders, areportfolio insurers, aided and abetted by stock in-dex arbitragers. Stock index arbitragers, recall,are the now familiar stars of the Triple-WitchingHour. (See "Thank God It's ... the SEC," Regu-lation, Sept/Oct 1986.) They buy and sell futuresand options contracts on stock indexes (con-tracts which generally are traded on the futuresexchanges, not on the stock exchanges), andsimultaneously sell and buy the stocks that makeup the indexes; the goal is to profit from pricingdiscrepancies between the futures market andthe stock market. Portfolio insurance, by con-

trast, is a sophisticated new hedging strategy; itsgoal is to protect a portfolio against loss in equityvalue while retaining some participation in priceappreciation.

A typical portfolio insurance program at-tempts to assure a minimum return over a speci-fied period of time, say 0 percent over threeyears. The "cost" of the insurance is theunderperformance of the portfolio in a risingmarket. The portfolio insurer (Leland O'BrienRubinstein Associates, Chase Investors Manage-ment Corporation Company, or Aetna Life andCasualty Company, for example) arranges andexecutes the program for a portfolio comprisingthe assets of many investors. Presently, there areabout a dozen portfolio insurers protecting over$50 billion in assets for pension funds, founda-tions, and endowments.

Portfolio insurance strategies can be (andwere originally) carried out directly in the cashmarkets. In this case, an initial hedge position iscreated by allocating part of the portfolio to eq-uities and part to Treasury bills or bonds. Thehedge is dynamically adjusted to increase protec-tion when the market is falling and to decreaseprotection when the market is rising. In a risingmarket, for example, the equity portion is in-creased while the Treasury bill or bond portionis simultaneously reduced.

Most current applications of portfolio insur-ance involve stock index futures, because of therelatively greater liquidity, speed of execution,and lower commissions associated with thesecontracts. Portfolio insurers generally take along position in stocks that is partially hedgedagainst the risk of a market decline by a shortposition in index futures. Index futures are soldin a declining market and bought (actually, theshort futures positions are offset by a long posi-tion) in a rising market.

The "meltdown" conjured by Phelan pro-ceeds as follows: An adverse event, such as aworsening of the trade deficit, causes the stockmarket to drop. Portfolio insurers begin sellingindex futures so as to reduce their equity expo-

sure and increase their hedged positions. Theprice of index futures falls, causing these con-tracts to sell at a discount relative to the basketof stocks underlying the index. When the dis-count reaches a critical value, stock index arbi-trage is triggered. Arbitragers step in to buy in-dex futures and sell the basket of stocksunderlying the index. These large block sales ofstocks depress stock prices, triggering portfolioinsurance programs again, prompting anotherwave of index-futures sales and price reductions.The cycle repeats itself, leading ultimately to acollapse of the stock market.

While dramatic, the meltdown scenariooverlooks a couple of important points. First,like most exercises in verbal dynamics, it ignoresthe feedback effects that tend to dampen explo-sive tendencies in markets. If stock prices weredeclining as a result of trading in index futures,for example, this would create buying interest instocks since their fundamental values would beunchanged. In addition, investors who buystocks after the market has fallen and sell stocksafter the market has risen-investors who can bethought of as "sellers" of portfolio insurance-would enter the market. For both reasons, thedecline in the stock market would tend to behalted. At the same time, the index arbitrage trig-gered by the availability of discounted index fu-tures would tend to bid up and stabilize futuresprices. As futures prices firmed up, there wouldbe fewer arbitrage opportunities, and thus lessarbitrage-related selling of stocks. To the extentthat the discount widened or persisted, portfolioinsurance strategies would become more costlyto implement. Portfolio insurers would tolerate alesser degree of accuracy in their hedges andwould limit their selling of index futures.

Second, if there is any merit to Phelan's ar-gument, the prospect of explosive growth in themarket is just as real as the prospect of a collapsein the market. The reason, of course, is that theevent that triggers the whole dynamic adjust-ment could be good news rather than bad news.As Securities and Exchange Commissioner Jo-seph Grundfest put it, "the market meltdownscenario can be translated into a story about themarket levitating itself-and that doesn't makemuch sense either."

In questioning the activities of portfolio in-surers and other program traders, Phelan seemsto have forgotten what the world was like in the1970s or even the early 1980s. In the days beforeindex futures and the sophisticated trading strat-

egies they have fostered, a declining market ledto stock sales, but in a far less orderly fashionthan is now possible. Portfolio managers simplybailed out of the equity market when losses ex-ceeded some critical amount. The portfolio in-surance strategies now in use permit institu-tional investors to carefully adjust equityexposure and avoid all-or-none decisions.

So much for the logic of portfolio insurance.What about the evidence of increased stock mar-ket volatility? Proponents of new federal regula-tion charge that since the development of indexfutures and index-related trading strategies, thestock market has experienced a sharp increase involatility. They point to days like September 11,1986, when the market fell 86 points-the larg-est daily price decline in history-or to February17, 1987, when the market increased 54 points-the largest daily increase in history. As noted inthe press, the drop in the market on that ill-fatedday, October 28, 1929, was 38 points, less thanthe size of six daily price declines in 1986.

Given the greatly increased value of thestock market of the 1980s, absolute changes indaily stock prices provide little useful informa-tion on market volatility. A 38-point drop in to-day's market would be the equivalent of a 2 to 3point drop in the 1920s, or a 13 to 14 point dropas recently as 1982. As some in the futures indus-try have been prompted to say, if market volatil-ity is a problem that can be measured by changesin the absolute level of daily stock prices, "theproblem could be solved by splitting the Dow 10for 1." When daily price changes are calculatedon a percentage basis, over the period 1940 to1987, only 3 of the 20 largest declines in the Dowoccurred since 1982, when index futures werefirst traded; 4 of the 20 largest increases oc-curred in this period.

Several careful empirical studies, conductedboth by the Securities and Exchange Commis-sion staff and by outside experts, reveal no in-crease in volatility related to the introduction ofindex futures. There is some evidence, in fact,that there may be less volatility. A recent studyby Dr. Franklin Edwards, Director of the Colum-bia Futures Center at Columbia University, findsthat the average daily percentage change (a mea-sure of interday volatility) in the Standard andPoor's 500 Index was significantly lower in theperiod 1982 to 1985 than in the period 1973 to1979. In an evaluation of the period 1980through 1986, G. J. Santoni, senior economist atthe Federal Reserve Bank of St. Louis, finds no

"DING DONG, THE WITCH ISDEAD!" -William J. Brodsky,President, Chicago MercantileExchange

On June 5, the Chicago MercantileExchange, the New York Stock Ex-change (NYSE), and the New YorkFutures Exchange (NYFE) jointlyannounced new settlement proce-dures for certain stock index fu-tures and options contracts, whichbecame effective for contracts ex-piring on Friday, June 19.

The new procedures are de-signed to ameliorate some of theprice volatility experienced in the"Triple-Witching Hour." Beforethe change, this was the final hourof trading on the third Friday ofeach quarter when stock index fu-tures, stock index options, and op-tions on individual stocks all ex-pired at the same time. Stock index

arbitragers generally close outtheir positions when the index con-tracts are settled, causing huge or-der imbalances for the stocks thatmake up the indexes. Before thechange in procedures, these orderimbalances could result in dra-matic price swings in the final min-utes of trading on the NYSE.

Under the new procedures, in-dex futures and options contractsare settled on the basis of openingstock prices on expiration Friday,rather than on closing prices, andtrading in these contracts termi-nates at the close of business thepreceding day. Stock orders relat-ing to expiring index contractsmust be received by 9:00 a.m. onexpiration Friday. The new proce-dures affect the Chicago Merc'sS&P 500 Index contracts, the mostactively traded of index futurescontracts, as well as certain con-

tracts traded on the NYFE and onthe NYSE. The NYSE and the Chi-cago Merc have also set up specialprocedures for disseminating in-formation on order imbalances for50 selected stocks.

The new procedures are basedon the belief that order imbalancescan be accommodated with mini-mal price effects in the morning.Before the opening, NYSE special-ists can disseminate informationabout order imbalances to attractorders on the opposite side. Theycan also delay the opening ofstocks when there are unusuallylarge imbalances.

It is too early to tell whether"the witch is dead." But therewere no significant price moves inthe stock market on the June 18close or on the open or close of themarket on June 19. The Septemberexpiration was quiet too.

significant change in interday volatility, but a sig-nificant decline in intraday volatility since 1982.As Santoni concludes, in an article published inthe St. Louis Federal Reserve Bank Review,"While closer scrutiny and regulation of tradingin stock index futures markets may be justifiedon other grounds, the evidence presented heresuggests that regulation based on the propositionthat it has increased price volatility in the spotmarket would be misdirected."

Notwithstanding all of the above, some in-vestors are getting edgy about a possible turn inthe market. That is not surprising. The stockmarket has, it would seem, defied all odds overthe past five years, breaking the 1,000 mark, the2,000 mark, and now the 2,500 mark. Any rea-sonable investor is likely to be having that nag-ging feeling: the bull market can't last forever.

The suggestion of a "market meltdown"seems well-timed to capitalize on this anxiety. Ifthe bears arrive, some fingers will undoubtedlybe pointed at Chicago, the center of trading inindex futures. But these new instruments did notcause the bull market nor will they end it.

Index futures provide a fast and efficientmeans of adjusting equity exposure and improv-ing the alignment of prices between the futuresmarket and the stock market. They have facili-tated the development of new risk managementstrategies that benefit institutional investors andthe millions of individuals they represent. Unfor-

tunately, these benefits are poorly understood bythe public and, if critics have their way, easilycould be lost in a push for regulation.

Investors, and the public generally, shouldbe wary of claims that new federal regulationsare necessary to prevent the destruction of thestock market-or any other market for that mat-ter. As Commodity Futures Trading Commis-sioner Robert Davis said recently, "Some com-plaints about portfolio insurance and otherindex-related trading strategies sound like a caseof sour grapes-they can often be traced to firmsthat are losing business because they don't havethe capability to compete in the market. Thisdoesn't mean you shouldn't ask questions, butit's our job to look through the smokescreen."

With the emergence of new financial instru-ments and new trading strategies, linking mar-kets both in the United States and abroad, com-petition between the exchanges and betweentraders is as intense as ever. The struggle be-tween the traditional market for securities, cen-tered in New York, and the new and rapidlygrowing market for financial futures and op-tions, centered in Chicago, is but one example ofthis competition. Let us hope that the regulatorscan resist the temptation to intervene in thisstruggle. The way in which the problems associ-ated with the Triple-Witching Hour were han-dled gives us reason to be optimistic.

Giving Back the Takings ClauseAfter decades of neglect, the Supreme Court in-fused new vitality into the "takings" clause of theFifth Amendment with two decisions handeddown in June. It is too soon to tell whether thesedecisions mark a major new beachhead in con-stitutionallaw. Nor is it clear whether these deci-sions should be credited as a victory for contem-porary market economists or for old-fashionedchampions of limited government. Still, theCourt's recent rulings hint strongly at a reemer-gence of constitutional checks on regulatory pol-icy making.

Prior to the 1930s, both state and federalgovernment regulation had to contend with a va-riety of constitutional obstacles, as the U.S. Su-preme Court broadly interpreted the Constitu-tion's restrictions on government powers.Regulations might be held to violate the constitu-tional prohibition against "impairing the obliga-tion of contracts," for example, or to transgress(in either direction) the boundaries between fed-eral interstate commerce powers and state po-lice powers. When no other constitutional clauseclearly applied, the Court was prepared on occa-sion to find economic restrictions in violation ofthe Fourteenth Amendment's guarantee againstdeprivation of liberty or property "without dueprocess of law." Most of this protective jurispru-dence was quickly and resolutely abandoned af-ter the Supreme Court's bruising confrontationwith the New Deal.

The takings clause of the Fifth Amendmentseemed to suffer the same fate as other constitu-tional checks on government. By its terms-" ... nor shall private property be taken for pub-lic use without just compensation"-the takingsclause seems to require compensation for anygovernmental takings of private property. Beforethe 1930s, the Supreme Court readilyacknowleged that the clause applied not only tooutright exercises of eminent domain powers,but also to regulatory restrictions on the free useof private property, "if," as Justice Holmes put itin a 1922 decision, "the regulation goes too far."Many regulations apparently did not go too far,however. The Court ruled that municipal zoningrestrictions (land use controls designed to pro-tect the character of particular neighborhoods atno cost to the residents) were legitimate, non-compensable regulatory actions. After the "con-stitutional revolution" of the late 1930s, theCourt was even less inclined to impose signifi-

cant compensation requirements when dealingwith regulatory restrictions, especially when therestrictions might be distinguished from com-plete, physical takings of property.

In four separate cases after 1980 where thisissue was pressed, the Court found proceduralgrounds to avoid setting forth a remedy forclaims of uncompensated regulatory takings. Aslate as March of 1987, when the Court finallygave a direct ruling on the merits of one suchclaim, it actually seemed to endorse a very broadscope for the imposition of regulatory restric-tions without compensation: in Keystone Bitumi-nous Coal Association v. DeBenedictis, the Courtupheld a Pennsylvania regulation restricting theamount of coal that could be extracted from par-ticular properties-effectively expropriatingsome 27 million tons of coal without compensa-tion. Ironically, this was virtually identical to theregulation that went "too far" for the 1922court.

Four justices, led by Chief Justice WilliamRehnquist, dissented from the ruling in Key-stone. The concerns of the dissenters found ma-jority support in the two rulings handed down inJune, which put a very different gloss on the reg-ulatory implications of the takings clause. InFirst English Evangelical Lutheran Church ofGlendale v. County of Los Angeles, a majority ofsix, including Justices William Brennan andThurgood Marshall, held that even a temporaryregulatory taking may justify damage claims bythe owner.

In this case, a church had purchased landfor a recreational center for handicapped chil-dren in 1957. When this center was destroyed bya flood in 1978, the county rezoned the land tobar any reconstruction. The church's suit fordamages was dismissed in the California courts,which held that compensation was limited tothose cases where the relevant regulatory restric-tion had already been judged by a court to be ataking and was still maintained without com-pensation by the government. The SupremeCourt, in an opinion by Chief Justice Rehnquist,found this approach to violate the takings clauseby leaving the initial property losses with own-ers. The Court insisted "that 'temporary' takingswhich, as here, deny a landowner all use of hisproperty, are not different in kind from perma-nent takings, for which the Constitution requirescompensation. "

The Court's decision in Nollan v. CaliforniaCoastal Commission, handed down less thanthree weeks later, may prove still more impor-

tant. The Nollan family had purchased a dilapi-dated beach bungalow and requested a permit todemolish the structure and build a new home.The Coastal Commission agreed to issue the per-mit only on the condition that the Nollans granta public easement over their land to the beach.The California Court of Appeals upheld this re-striction on the permit, but the U.S. SupremeCourt, in a 5 to 4 decision, ruled that it violatedthe takings clause. Justice Antonin Scalia's ma-jority opinion described the Coastal Commis-sion's conditional permit as "an out-and-out planof extortion."

The Court acknowledged that the CoastalCommission might properly condition its con-struction permit on compliance with some rele-vant zoning conc·ern, such as maintaining an un-obstructed public view of the beach, whichmight otherwise justify a total prohibition onconstruction. But the commission could not im-pose a conditional permit where "the condition... utterly fails to further the end advanced asthe justification for the prohibition." Increasingpublic access to the beach might be a worth-while goal, the Court concluded, but if the state"wants an easement across the Nollan's prop-erty, it must pay for it."

The Nollan decision is consistent with thebroad trend of post-New Deal takings cases indeferring to governmental conceptions of "pub-lic use." The Court did not question California'spower to force the Nollans to make an easementthrough their property, so long as the state wasprepared to pay "just compensation" for its de-mands-just as the Court did not question theauthority of Los Angeles County to rezone thechurch's property in First English EvangelicalChurch. But Nollan may still have far-reachingimplications for land use regulation, insofar as itsuggests a more limited constitutional scope foruncompensated regulatory restrictions.

Many municipalities and counties aroundthe country now routinely adopt "inclusionaryzoning" measures. Instead of merely seeking toexclude some property uses as neighborhooddisturbances, such measures force developers topursue particular, governmentally favored uses.For example, developers have been required toconstruct a certain amount of low-cost housingin order to receive building permits. Many Cali-fornia cities also condition their approval of con-struction plans on developer contributions topublic day care, mass transit, and arts projects. Ifthe courts follow the logic of the Nollan decision,

they may come to regard more and more of thesemeasures as "out-and-out extortion"-so long asgovernments offer no compensation for the costsinvolved. And under First English, governmentsmay have to pay for the delays imposed by theirefforts to negotiate such schemes, even if theysubsequently give them up as too expensive.

It is far too early to tell where the boundarybetween compensable and non-compensableregulatory impositions will be drawn. But itseems clear that whether or not the Rehnquistcourt finds absolute constitutional limits on gov-ernment powers, it will sometimes use thetakings clause to present governments with a billfor damages when their actions tread too heavilyon private property rights.

This approach may not satisfy rigorous nine-teenth-century champions of property rights andlimited government-nor their contemporarylibertarian followers. But it may be a positionthat is more defensible-and perhaps more sus-tainable-for unelected judges in an era whenpublic opinion is sharply divided on the properrole for government. Economists and other ana-lysts concerned about the costs of regulationshould be heartened by a constitutional doctrinethat forces more of these costs into the open.Perhaps when more of these costs are laid di-rectly on taxpayers, we will see a more settledand encompassing public consensus developabout what governments should and should nottry to control.

An Environment of Riskand Uncertainty

In a recent public address, President Reaganurged the Senate to confirm his Supreme Courtnominee, Judge Robert Bork. He correctly notedthat, "As a member of [the U.S. Court of Ap-peals], Judge Bork has written more than 100majority opinions and joined in another 300. TheSupreme Court has never reversed a single oneof these 400 opinions."

Ironically, just 15 days before the Presidentspoke, Judge Bork reversed himself. Writing fora unanimous eleven-judge D.C. Circuit Court ofAppeals (in Natural Resources Defense Council v.U.S. Environmental Protection Agency), Borkoverturned his own majority opinion, issued just

eight months earlier by a three-judge panel of thesame court. Bork's second opinion in this caseappears to undercut longstanding decision-mak-ing practices at the Environmental ProtectionAgency and other regulatory agencies, and maymark an important watershed in health andsafety law.

The court vacated the EPA's 1985 decisionto withdraw a proposed emissions standard forvinyl chloride. The proposal was made in 1977,and would have tightened an emissions standardissued the previous year. In granting the NaturalResources Defense Council's (NRDC) petitionfor review and remanding the decision to theEPA, the court found that the Clean Air Act re-quired the EPAto make a determination that theemissions levels under the 1976 rule were"safe." Its instructions to the agency on how todecide what is safe and when to consider costwill affect health and safety regulation through-out the federal government.

The debate in health and safety regulationhas long been between those who balance safety(or risk) against cost and those who refuse to doso. It is sometimes possible to reconcile thesetwo positions, as when agencies regulate chemi-cals that appear to have thresholds of toxicity. Inthese cases, zero risk can be attained at less thaninfinite cost by setting acceptable exposure atthe "no effects" level. These cases are rare, how-ever. Modern techniques can detect the pres-ence of chemicals at the parts-per-billion level orlower. Cancer is the most prominent (if not themost important) health effect of concern, andrisk assessment methods predict non-zero can-cer risk for any non-zero dose. For most chemi-cals of interest, therefore, no-effects thresholdshave disappeared as an obvious point of compro-mise. As noted in the first sentence of Bork'sopinion:

Current scientific knowledge does not per-mit a finding that there is a completely safelevel of human exposure to carcinogenicagents.

The risk-cost balancers and the zero-riskershave become increasingly polarized in recentyears, leaving many legislators and policy mak-ers in an uncomfortable position. Congress haswritten environmental statutes that explicitly re-quire risk-cost balancing, notably the Toxic Sub-stances Control Act of 1976, and the Federal In-secticide, Fungicide, and Rodenticide Act of1978 (FIFRA). It has also written zero-risk stat-

utes, such as the infamous Delaney clause of theFood, Drug, and Cosmetic Act, which applies tosome pesticides as well as to food additives. Butin most cases, Congress has ducked the question,telling OSHA to ensure that the workplace is"safe" and the EPA to "protect human healthand the environment." The regulators (and ulti-mately the judges) are left to figure out what,precisely, Congress means.

In the case of the vinyl chloride emissionsstandard, the EPAconcluded that, in passing theClean Air Act, Congress could not have meantfor hazardous emissions to be totally banned.The court agreed:

The EPA has determined that a zero-emis-sions standard for non-threshold pollutantswould result in the elimination of such ac-tivities as "the generation of electricityfrom either coal-burning or nuclear energy;the manufacturing of steel; the mining,smelting, or refining of virtually any min-eral; the manufacture of synthetic organicchemicals; and the refining, storage, or dis-pensing of any petroleum product." It issimply not possible that Congress intendedsuch havoc on the American economy andnot a single representative or senator men-tioned the fact.

If agencies are to avoid wreaking havoc,however, they must find some "stopping point"at which they can rationalize a non-zero stan-dard. They cannot do this if they focus exclu-sively on the objective of safety.

One common stopping point is "economicfeasibility," which is code for clemency. It usu-ally means that a firm or industry may bebrought to the brink of bankruptcy, but shouldnot actually be shut down. Some health andsafety statutes make economic feasibility an ex-plicit consideration for rulemaking; in othercases agencies apply it as a form of prosecutorialdiscretion. It is politically popular because it em-bodies the principle that shareholders' equitycan be appropriated in the pursuit of environ-mental goals, but not jobs-at least not in an ob-vious way. Many economists criticize this proce-dure because it rewards inefficient firms andpunishes efficient ones, and because it is used torationalize goals that appear "affordable" ratherthan those that are worthwhile. Economistswithin agencies, however, tend to be less critical.Economic feasibility is often the only form ofeconomic analysis that agencies are able-orwilling-to use.

MR .WALS~SI15 OOWNTOA QUlITEVEN~ DINNER ~A~ rw AT M~CE. Ht~ N11JE FA1rY ST'fAJ(--mE eROCC.OUWIT11rr;lIIODE PfSIOOE- M WKESSED ~HED~1CES Wl'1l1A RlJND(f amR-llf 1l1REE STRIPS OFBtl.CN- TIfE Ulll.E~N &Wl a= SACCHbRlN-1HE Q.JPOF STEt»'11f\li COFFEE. MR. WALSH SUOCENLY C<MS lD ASfAmlNG REALIZATION: HISWIFE 15~1N6m KIU- HIM.

Another common stopping point is techno-logical feasibility. At the EPAthis takes the formof "technology-based" standards. Agency engi-neers decide what the "best" technology is for aparticular industry, and require all firms to in-stall it. Sometimes this is explicitly required bystatute; in other cases the EPA has used its ad-ministrative discretion to propose technology-based standards.

In the wake of the court's decision, these tra-ditional stopping points will have to be re-thought. The vinyl chloride emissions standardbefore the court was technology-based. Section112 of the Clean Air Act requires the EPAAdmin-istrator to set standards "at the level which in hisjudgment provides an ample margin of safety toprotect public health." The Administrator didnot believe that a total ban on vinyl chlorideemissions was justifiable, and interpreted hismandate to require the best available technol-ogy. The 1976 vinyl chloride standard requires

firms to use what the EPA thought at the timewas the "best" technology. In 1985, the EPAwithdrew its proposed revision because it stillbelieved the 1976 technology was the best avail-able. Not good enough, said the court. You mustdecide what is "safe," and you must do so with-out regard to cost or feasibility. In Bork's words:

This determination must be based solelyupon the risk to health. The Administratorcannot under any circumstances considercost and technological feasibility at thisstage of the analysis. The latter factors haveno relevance to the preliminary determina-tion of what is safe. Of course, if the Admin-istrator cannot find that there is an accept-able risk at any level, then the Admin-istrator must set the level at zero.

What risk is acceptable if the costs of elimi-nating it cannot be considered? The only rationalanswer seems to be no risk at all. The court doesnot necessarily agree:

... the Administrator's decision must bebased upon an expert judgment with regardto the level of emissions that will result in an"acceptable" risk .... [He] must determinewhat inferences should be drawn from avail-able scientific data and decide what risks areacceptable in the world in which we live.('There are many activities that we engagein every day-such as driving a car or evenbreathing city air-that entail some risk ofaccident or material health impairment;nevertheless, few people would considerthese activities 'unsafe.' ")

Ralph Nader does and so do many others.Drawing unambiguous inferences about whatpeople regard as safe will be no small task.

Once a "safe" level has been determined, ac-cording to the court, the Administrator must de-cide what "margin of safety" to apply. This, thecourt reasons, is intended to take account of sci-entific uncertainty in the degree of risk. Here theAdministrator has a great deal of discretion, andmay consider costs, technological feasibility, andwhatever else he deems relevant to setting a rea-sonable margin of safety.

In Brief-NAAGing the Airlines. The Na-tional Association of AttorneysGeneral (NAAG) recently gave theJustice Department's Antitrust Di-vision some competition by issuingits own merger guidelines thatcontrasted sharply with the federalmerger guidelines. Now NAAG isplanning to compete with the fed-eral Department of Transportation(DOT). It has formed a task forceof 20 state attorneys general whohave written standards for regulat-ing airfare advertisements.

The proposed standards includerequirements to:• enlarge the print size of the re-strictions in advertisements, andplace the restrictions in a specialbox or in the middle or top of theadvertisements;• advertise round-trip prices (notone-way prices for half the round-trip price, unless one-way ticketscan be purchased at the prices that

are listed); and• make a reasonable number ofseats available at the advertisedprices.NAAG will vote on the revisedguidelines in December.

While it is not clear that the re-vised guidelines would be en-forced even if approved (the Air-line Deregulation Act of 1978 givesthe DOT jurisdiction over airlineadvertising), NAAG's action putsall federal regulatory agencies onnotice. If the agencies show re-straint in the face of a demand forregulation, NAAG stands ready tostep in and increase the supply ofregulation.

Reel Colors. The U.S. CopyrightOffice's June ruling that computercolorizations of black-and-whitefilms are entitled to copyright pro-tection has created a real stir in themovie industry.

According to the rules, colorizedversions of old movies have protec-tion for 75 years as "derivativeworks." Studios that producecolorized films (including Walt

Disney Productions, 20th CenturyFox, and Turner Broadcast Sys-tems) will be able both to sell thefilms to broadcast and cable televi-sion stations and videotape distrib-utors, and to collect damages ifother firms copy their products.Copyrights will cover only thecolor selections added to films, notthe films themselves; films will re-main in the public domain if theiroriginal copyrights have expired.

In making its decision, the Copy-right Office stressed the creativerole of human operators in gener-ating colorized films, noting thatthere are more than 16 milliontints and tones that can be selectedfor computer colorizing. This em-phasis hints at a possible futurepolicy issue: Are the protections ofintellectual property law availableto the intellectual product of com-puters? Someday soon, computersmight be able to produce a taste-fully colorized movie without rely-ing much on the color judgmentsof humans.

Members of the Directors Guildof America have voiced outrage at

~." ',' •..•.•• ,••,".,., ''''. \~A\lE.S 7-11(,

© 1987 Newspaper Enterprise Association, Inc.

It is difficult to say where this discretion be-gins. "Risk" is generally used to refer to an un-biased probabilistic assessment of the harm thatis expected to result. "Uncertainty" is somethingdifferent in the court's view. The most plausibledistinction is that it is some measure of the vari-ance in the estimate of risk.

This is an important distinction, since muchof what passes for environmental risk is reallyuncertainty. Many chemicals (although not vinylchloride) are classified as carcinogens on the ba-sis of animal evidence, and the projected risk es-

timates to humans are highly speculative. At ev-ery stage of a risk assessment, regulatory agen-cies are prone to incorporate worst-case scenar-ios, "conservative" assumptions, and up-per-bound estimates. (See Albert L. Nichols andRichard J. Zeckhauser, "The Perils of Pru-dence," Regulation, Nov/Dec 1986.) All of thesetechniques are intended to create a margin ofsafety, and are therefore matters of discretion.They are, in the terms made popular by the Na-tional Academy of Sciences, properly regardedas part of the risk management process, and not

the Copyright Office's decision.They claim an inalienable "moralright" to control the content oftheir artistic product. Two bills arenow pending in Congress thatwould give directors and screen-writers control over filmcolorizing, regardless of who actu-ally owns the copyright.

Regardless of how these moralrights and intellectual propertyrights are sorted out, colorizationis a technology that enjoys strongpublic demand. In less than a yearas a color movie, "Captain Blood"grossed $800,000, four times theamount it grossed in its history as ablack and white movie.

Chernobyl Fallout. One of the fac-tors that makes environmental"risk-management" difficult is thegap between the experts' judgmentof relative risks and the public'sperception of those same risks.Consider the accident at ThreeMile Island in 1979. This eventgreatly aggravated public appre-hensions about nuclear power. Inits aftermath, however, the ex-perts reduced their estimates ofthe riskiness of nuclear powerplants. By now, public concern hasvirtually ruled out nuclear poweras a viable option for making elec-tricity, despite the heavy toll of hu-man lives associated with the pro-duction and use of one of itsprincipal alternatives, coal.

Opponents of nuclear powermight feel vindicated by the nu-clear accident at Chernobyl. Onceagain, though, there is expert opin-ion on the other side. Harvard pro-fessor Richard Wilson, in a recentarticle in Science magazine, esti-

mates that a nuclear reactor wouldhave to go through a Chernobyl-sized meltdown every year to beapproximately as hazardous as anequivalent-sized coal-burningpower plant.

Vying Over Spilt Milk. The NewYork State Legislature, in one of itsfinal actions before summer re-cess, passed a bill ending thestate's milk marketing laws that foryears had restricted entry into themilk distribution market. The leg-islature was following the lead of aFederal court, which in Januaryhad set aside the state law to allowa New Jersey processor, FarmlandDairies, to market milk in NewYork City. (See "The Milking ofNew York City," Regulation, 1987Number 1, and Letters in thisissue.)

Farmland's entry into the mar-ket has sparked intense price com-petition. Between January andJuly, the retail price of whole milkin the New York Metropolitan areadropped by almost 10 percent.Though milk prices are normallysomewhat lower in the summer,this drop is largely attributable tothe collapse of the distributors'cartel in the city.

But consumers may not get toenjoy lower milk prices for long.On July 10, the Regional Cooper-ative Marketing Agency (RCMA),which includes farmers in elevenstates in the Northeast, includingNew York and New Jersey, an-nounced a new minimum price forraw milk. The U.S. Department ofAgriculture already sets minimumprices for wholesale fresh milk un-der its marketing order program.

The RCMA has antitrust immunityfor setting "over-order" prices. Itsminimum price is $1.25 per gallon,2 cents higher than the federalminimum price.

It remains to be seen whetherthe RCMA can make its higherprice stick; even cartels with anti-trust immunity face serious en-forcement problems when theycannot invoke the coercive powersof federal or state governments.

Food for Health. On August 4, theFood and Drug Administration(FDA) proposed new regulationsfor food labeling that would allowfood manufacturers to advertisethe beneficial health effects oftheir products (52 Federal Register28843). The proposed rule wouldpermit messages that are truthfuland substantiated by clear scien-tific evidence, but does not imposea "consensus" requirement-anagreement of experts for eachhealth claim-as stipulated in anearlier draft. The proposal wouldset up a committee of experts fromthe FDA, the Department of Agri-culture, the Public Health Service,and the Federal Trade Commissionto develop exemplary health mes-sages to guide advertisers; whetheradvertisers would be bound by thecommittee's suggestions is stillunclear.

Formerly, the FDA fought tokeep health claims out of food la-beling. (See "A Food By Any OtherName," Regulation, 1987 Number1.) Kellogg's All-Bran advertise-ments challenged this policy in1984, initiating a debate that stillcontinues. Stand by for the FDA'sfinal decision.

part of the risk assessment process.We are left with an absolute mandate to

achieve a relatively low risk with a discretionarydegree of certainty. The court can be forgiven fornot attempting to make a silk purse out of asow's ear. This confused paradigm may be a rea-sonable reading of an unreasonable Congress'sintent. But with what result? Regulators willhave to identify some acceptable level of safety;the quest for consistency, economy, and cer-tainty across agencies will create strong pres-sures for a single number that characterizes whatthat level is.

In response to the court's decision, environ-mentalists can be expected to advocate a low, ifnot precisely zero, level of risk as safe. A leadingcandidate is one in a million. That number, 10-6,

has been emerging as a sort of consensus view ofthe de minimis level of risk for all sorts of envi-ronmental hazards. Its name comes from a legalprecept: de minimis non curat lex, or "the lawdoes not concern itself with trifles." The idea isthat some disputes are technically actionableunder the law, but are so insignificant that nopublic policy argument can be advanced for con-suming the resources of the legal system in theirresolution.

De minimis risk has lately become a flagaround which would-be moderates-those whoare reluctant to embrace either the zero-risk orthe cost-risk balancing view-are rallying. Sim-ply by defining zero as anything less than 10-6, athreshold is restored that quantitative risk-assess-ment models had taken away. Earlier this year,the National Academy of Sciences sent the EPAareport on FIFRA, urging the EPA and Congressto adopt a uniform de minimis threshold of 10-6

for registered pesticides, in place of the currentstatutory mandate to balance benefits and costs.

It seems clear, however, that the court didnot have a de minimis risk in mind. De minimisdefines a floor below which government agen-cies cannot reasonably go even when the Con-gress mandates zero risk. The court, on the otherhand, seems to contemplate a ceiling abovewhich things are manifestly unsafe and belowwhich the law and the agency may still be con-cerned. Between the two, one can imagine a"zone of reasonableness" within which agenciesmay exercise discretion.

Apaper published this year describes exactlysuch a zone of reasonableness. (Curtis C. Travis,Samantha A. Richter, Edmund A. C. Crouch,Richard Wilson, and Ernest D. Klema, "Cancer

Risk Management: A Review of 132 Federal Reg-ulatory Decisions," in Environmental Scienceand Technology, Vol. 21, No.5, 1987.) The au-thors attempt to measure the levels of risk thatcause agencies to act. They coin a new term todemark the ceiling: de manifestis risk, "literally arisk of obvious or evident concern, ... one that isinstantly recognized by a person of ordinaryintelligence." For small populations, they findthe de manifestis level to be about 4 times 10-3, or4 in 1,000. For large populations at risk (wherelarge is approximately the entire U.S. popula-tion), it drops to 3 in 10,000.

The authors report that below the deminimis level of risk (which they also find to varywith the size of the exposed population), agen-cies almost never act. Above the de manifestislevel, they always act. In between those levels,the decision to regulate is triggered by a thresh-old of roughly $2 million per life saved-an indi-cation that cost balancing is a common practiceeven at agencies that typically disavow it.

This is an observation about what agenciesdo, as distinct from a normative statement ofwhat they should do, or a legal analysis of whatthey are obliged to do. If the courts find such apattern of decision making acceptable, then theBork decision may not require a total abandon-ment of established decision-making habits.

Note, however, that the risk estimates usedin this analysis were those presented by the agen-cies making the decisions. Some were unbiasedestimates; others were exaggerated to incorpo-rate large margins of safety. Even if a single num-ber, or a zone of reasonableness, emerges fromthe court's decision, regulatory policy will notnecessarily become more uniform or sensible.The technocrats who produce risk estimates willstill be prone to exaggerate risks. As long as thedegree of uncertainty, and the degree of exagger-ation, vary so greatly across risks and agencies, apolicy of uniformity will produce little morethan the illusion of uniformity.

If the EPAAdministrator could cleanly sepa-rate estimates of risk from associated estimatesof uncertainty, he might find that the court's de-cision left him far more discretion than it ap-pears. Judging by past experience, however, riskassessments will be fully cooked by the time theyreach the Administrator's desk. Between the law-yers and the technocrats, it is unlikely that muchdiscretion will be left to the appointees who arecharged with making the decisions.